Saving the Ocean with Neural Networks

Inspiration

This was inspired partially by the theme of the hackathon and partially by my team's love for AI and Neural Networks.

What it does

This software can predict the location and density of plastic trash in the ocean.

How we built it

We built it from the ground up in java and trained it with a large supply of data using genetic algorithms.

Challenges we ran into

It was challenging to figure out the loss function necessary to make the genetic algorithm work properly.

Accomplishments that we're proud of

We are proud of the real world application of the product, and how it can make it easier for us to clean up plastic where it counts.

What we learned

Machine learning is powerful but not necessarily perfect- especially if the data isn't. The data had some minor flaws concerning incomplete data, so we had to make some estimates to allow the algorithm to work correctly.

What's next for Saving the Ocean with Neural Networks

Using more advanced machine learning models to make the product more accurate, and adding a visualization tool that plots the algorithm's results to a real-world map.

Built With

• java